Method and device for determining traffic flow and electronic equipment

Abstract

The invention relates to a traffic flow determination method and device and electronic equipment, and relates to the technical field of vehicles, and the method comprises the steps: obtaining a plurality of traffic flow data of a target region in a first time period; the traffic flow data correspond to different data sources; carrying out feature fusion on the plurality of traffic flow data to obtain a traffic feature data set of the target area, and carrying out time sequence analysis on the traffic flow data set to obtain traffic flow change information of the target area; determining predicted traffic flow data of the target area in the second time period based on the change information of the traffic flow of the target area; the second time period is after the first time period. Therefore, the traffic pressure can be reduced and the traffic efficiency can be improved by predicting the traffic flow.

Description

Technical Field

[0001] The present application relates to the field of vehicle technology, in particular to the field of traffic flow determination technology, and specifically to a method, device and electronic equipment for determining traffic flow. Background Art

[0002] In today's society, with the acceleration of urbanization and the continued growth of the vehicle population, traffic congestion has become a major problem facing most cities around the world. This not only affects people's daily travel efficiency but also has a negative impact on urban economic development and residents' quality of life. To alleviate traffic pressure and improve road efficiency, scholars and engineers around the world have conducted extensive research, especially in the development of traffic flow prediction and intelligent transportation systems.

[0003] In one existing technology, data such as traffic flow parameters, weather condition parameters, and emergency event parameters are collected in real time through the 5G network. The data of traffic flow parameters, weather condition parameters, and emergency event parameters are fused and processed, and the comprehensive traffic status is determined based on the fusion algorithm. A control strategy is generated and applied based on the traffic status, and traffic lights and electronic signs are controlled through 5G terminals to dynamically adjust traffic flow and guide vehicles to avoid congested areas. One existing technology uses a graph convolutional network (GCN) and adaptive spatial weight assignment (ASWA) mechanism to capture spatial dependencies in traffic networks. It combines a long short-term memory network (LSTM) with an attention mechanism to capture both long-term and short-term dependencies in time series. It uses feature embedding, dynamic impact assessment, and fusion prediction to capture the dynamic changes of external factors and their impact on traffic flow. The trained model is deployed in actual traffic management systems to provide real-time traffic flow predictions.

[0004] Therefore, how to accurately predict the traffic flow in the target area in the future and improve travel efficiency has become a technical problem that needs to be solved urgently. Summary of the Invention

[0005] This application provides a method, device, and electronic device for determining traffic flow to at least address the related art issue of how to accurately predict traffic flow in a target area over a period of time in the future and improve travel efficiency. The technical solution of this application is as follows: According to a first aspect provided by the present application, a method for determining traffic flow is provided, the method comprising: obtaining multiple traffic flow data of a target area within a first time period; the multiple traffic flow data correspond to different data sources; performing feature fusion on the multiple traffic flow data to obtain a traffic feature dataset of the target area, and performing time series analysis on the traffic flow dataset to obtain change information of the traffic flow of the target area; based on the change information of the traffic flow of the target area, determining predicted traffic flow data of the target area within a second time period; the second time period is located after the first time period.

[0006] Based on the above technical means, this application obtains multiple traffic flow data of the target area within a first time period, performs feature fusion on the multiple traffic flow data, and obtains a traffic feature dataset of the target area. A time series analysis is performed on the traffic feature dataset to obtain traffic flow change information of the target area, and based on this change information, the predicted traffic flow data of the target area within a second time period is determined. By fusing data from multiple data sources and the different feature data corresponding to the data sources, the changing trend of traffic flow under the influence of different feature data is analyzed. The traffic flow data of the target area in the future can be accurately predicted, thereby improving travel efficiency.

[0007] In one possible approach, a traffic feature data set includes linear feature data and nonlinear feature data; a time series analysis is performed on the traffic flow data set to obtain change information of the traffic flow in a target area, including: inputting the linear feature data into a linear data analysis model, analyzing the linear feature data through the linear data analysis model, and obtaining first change information of the traffic flow in the target area; inputting the nonlinear feature data into a nonlinear data analysis model, analyzing the nonlinear feature data through the nonlinear data analysis model, and obtaining second change information of the traffic flow in the target area; and performing data fusion on the first change information and the second change information to obtain change information of the traffic flow in the target area.

[0008] Based on the above technical means, this application uses a linear data analysis model to analyze linear feature data to obtain first information on changes in traffic flow in the target area, and uses a nonlinear data analysis model to analyze nonlinear feature data to obtain second information on changes in traffic flow in the target area. By analyzing these two data types, we comprehensively analyze the changes in traffic flow in the target area as affected by both linear and nonlinear data, providing reliable technical support for subsequent prediction of traffic flow data.

[0009] In one possible approach, based on the traffic flow change information of the target area, the predicted traffic flow data of the target area in the second time period is determined, including: inputting the traffic flow change information of the target area into the traffic flow prediction model, and the predicted traffic flow data of the target area in the second time period.

[0010] Based on the above technical means, this application inputs the traffic flow change information of the target area into the traffic flow prediction model, determines the predicted traffic flow data of the target area in the second time period, and predicts the traffic flow data of the target area in the second time period through the traffic flow prediction model, thereby improving the precision and accuracy of the predicted data.

[0011] In one possible approach, the traffic flow prediction model is determined by: inputting target historical traffic flow data into an initial traffic flow prediction model to obtain a prediction result corresponding to the target historical traffic flow data; when the deviation between the prediction result corresponding to the target historical traffic flow data and the actual result corresponding to the target historical traffic flow data is greater than a preset threshold, iteratively training the initial traffic flow prediction model until the difference between the prediction result of the trained traffic flow prediction model and the actual result is less than or equal to the preset threshold, and the trained traffic flow prediction model is used as the traffic flow prediction model.

[0012] Based on the above technical means, this application trains an initial traffic flow prediction model. If the difference between the predicted results of the trained traffic flow prediction model and the actual results is less than or equal to a preset threshold, the trained traffic flow prediction model is used as the traffic flow prediction model. This provides technical support for subsequent traffic flow predictions, making the prediction results more accurate.

[0013] In one possible approach, feature fusion is performed on multiple traffic flow data to obtain a traffic feature data set of a target area, including: extracting features from the multiple traffic flow data using a feature extraction function to obtain a feature parameter set corresponding to each traffic flow data; the feature parameter set includes traffic flow feature data at multiple moments in a first time period; and the traffic feature data at the same moment in the feature parameter set corresponding to the multiple traffic flow data is fused to obtain a traffic feature data set of the target area.

[0014] Based on the above technical means, this application uses a feature extraction function to extract features from multiple traffic flow data sets, obtains a set of characteristic parameters corresponding to each traffic flow data set, and fuses the traffic feature data at the same time within the feature parameter sets corresponding to the multiple traffic flow data sets to obtain a traffic feature dataset for the target area. By extracting and fusing features from traffic flow data to obtain a traffic feature dataset, the subsequent data analysis process is more convenient and efficient.

[0015] In one possible approach, traffic feature data at the same time in a feature parameter set corresponding to multiple traffic flow data are fused to obtain a traffic feature data set of a target area; based on the weight corresponding to each data source, traffic feature data at the same time in a feature parameter set corresponding to multiple traffic flow data are fused to obtain a first feature data set of the target area; the first feature data set includes fused feature data at each of the multiple moments, and a convolution operation is performed on the fused feature data at the multiple moments included in the first feature data set to obtain a traffic feature data set of the target area.

[0016] Using the aforementioned technical approach, traffic characteristic data from the same moment in time within a set of characteristic parameters corresponding to multiple traffic flow data points are fused based on the weights assigned to each data source to generate a first characteristic dataset for the target area. A convolution operation is then performed on the fused characteristic data from multiple moments within the first characteristic dataset to generate a traffic characteristic dataset for the target area. This fusion of characteristic data provides a data foundation for subsequent model analysis of traffic flow changes, simplifying the analysis process.

[0017] In one possible approach, multiple traffic flow data of a target area within a first time period are obtained, including: obtaining traffic information of each data source within the first time period; the traffic information includes traffic data at multiple moments, and the traffic data at each moment includes multiple different types of traffic data; for the traffic data at each moment in each data source, the multiple different types of traffic data included at each moment are integrated to obtain traffic flow data corresponding to the data source, so as to obtain multiple traffic flow data.

[0018] Based on the above technical means, this application integrates multiple different types of traffic data at each moment in each data source to obtain multiple traffic flow data, providing data support for subsequent data analysis and simplifying subsequent data analysis operations.

[0019] In one possible approach, multiple different types of traffic data included at a moment are fused to obtain traffic flow data corresponding to the data source, including: determining the weight of each type of traffic data; and based on the weight of each type of traffic data, fusing multiple different types of traffic data to obtain traffic flow data corresponding to the data source.

[0020] Based on the above technical means, this application determines the weight of each type of traffic data and, based on the weight of each type of traffic data, fuses multiple different types of traffic data to obtain the traffic flow data corresponding to the data source. By fusing traffic data based on the weights of multiple different types of traffic data, the form of traffic flow data in the data source is simplified, providing a data foundation for subsequent data processing.

[0021] In one possible embodiment, the method further includes: performing route planning based on traffic flow forecast data of the target area within the second time period, and determining a target driving route within the second time period; the target driving route includes a route with the shortest driving time and / or a route with the least vehicles among multiple routes within the target area; and sending the target driving route to the target vehicle.

[0022] Based on the above technical means, the present application performs route planning based on the predicted traffic flow data for the target area during the second time period, determines the target driving route during the second time period, and transmits the target driving route to the target vehicle. Route planning based on traffic flow prediction data improves travel efficiency and reduces traffic pressure.

[0023] In one possible embodiment, the method further includes: determining the signal light control information of the target area within the second time period based on the predicted traffic flow data of the target area within the second time period; the signal light control information is used to control the display duration of the red and green lights of the signal lights in the target area.

[0024] According to the above technical means, the present application determines the signal light control information of the target area in the second time period based on the predicted traffic flow data of the target area in the second time period. Based on the predicted traffic flow data, the traffic pressure in the target area is reduced from the perspective of signal light control.

[0025] According to a second aspect of the present application, the present application provides a traffic flow determination device, the device comprising: An acquisition module, configured to acquire a plurality of traffic flow data of a target area within a first time period; The determination module is used to perform feature fusion on multiple traffic flow data to obtain a traffic feature dataset of the target area, and perform time series analysis on the traffic flow dataset to obtain traffic flow change information of the target area; The determination module is further configured to determine traffic flow prediction data of the target area within a second time period based on the change information of the traffic flow in the target area.

[0026] In one possible approach, a module is determined, wherein a traffic feature data set includes linear feature data and nonlinear feature data; a time series analysis is performed on the traffic flow data set to obtain change information of the traffic flow in the target area, including: inputting the linear feature data into a linear data analysis model, analyzing the linear feature data through the linear data analysis model, and obtaining first change information of the traffic flow in the target area; inputting the nonlinear feature data into a nonlinear data analysis model, analyzing the nonlinear feature data through the nonlinear data analysis model, and obtaining second change information of the traffic flow in the target area; and performing data fusion on the first change information and the second change information to obtain change information of the traffic flow in the target area.

[0027] In one possible approach, the determination module determines the predicted traffic flow data of the target area in the second time period based on the change information of the traffic flow in the target area, including: inputting the change information of the traffic flow in the target area into the traffic flow prediction model, and the predicted traffic flow data of the target area in the second time period.

[0028] In one possible approach, the traffic flow prediction model of the determination module is determined in the following manner: target historical traffic flow data is input into the initial traffic flow prediction model to obtain a prediction result corresponding to the target historical traffic flow data; when the deviation between the prediction result corresponding to the target historical traffic flow data and the actual result corresponding to the target historical traffic flow data is greater than a preset threshold, the initial traffic flow prediction model is iteratively trained until the difference between the prediction result of the trained traffic flow prediction model and the actual result is less than or equal to the preset threshold, and the trained traffic flow prediction model is used as the traffic flow prediction model.

[0029] In one possible approach, a module is determined to perform feature fusion on multiple traffic flow data to obtain a traffic feature data set of a target area, including: extracting features from the multiple traffic flow data using a feature extraction function to obtain a feature parameter set corresponding to each traffic flow data; the feature parameter set includes traffic flow feature data at multiple moments in a first time period; and fusing the traffic feature data at the same moment in the feature parameter set corresponding to the multiple traffic flow data to obtain a traffic feature data set of the target area.

[0030] In one possible approach, a determination module is configured to fuse traffic feature data at the same time in a feature parameter set corresponding to multiple traffic flow data to obtain a traffic feature data set of a target area; based on a weight corresponding to each data source, traffic feature data at the same time in a feature parameter set corresponding to multiple traffic flow data are fused to obtain a first feature data set of the target area; the first feature data set includes fused feature data at each of the multiple moments, and a convolution operation is performed on the fused feature data at the multiple moments included in the first feature data set to obtain a traffic feature data set of the target area.

[0031] In one possible approach, a determination module obtains multiple traffic flow data of a target area within a first time period, including: obtaining traffic information of each data source within the first time period; the traffic information includes traffic data at multiple moments, and the traffic data at each moment includes multiple different types of traffic data; for the traffic data at each moment in each data source, the multiple different types of traffic data included at each moment are integrated to obtain traffic flow data corresponding to the data source, so as to obtain multiple traffic flow data.

[0032] In one possible approach, a determination module is configured to fuse multiple different types of traffic data at a given moment to obtain traffic flow data corresponding to the data source, including: determining a weight for each type of traffic data; and fusing multiple different types of traffic data based on the weight for each type of traffic data to obtain traffic flow data corresponding to the data source.

[0033] In one possible manner, the determination module and the method further include: performing route planning based on traffic flow prediction data of the target area within the second time period, and determining a target driving route within the second time period; the target driving route includes a route with the shortest driving time and / or a route with the least vehicles among multiple routes within the target area; and sending the target driving route to the target vehicle.

[0034] In one possible manner, the determination module and the method further include: determining the signal light control information of the target area within the second time period based on the traffic flow of the target area within the second time period; the signal light control information is used to control the display duration of the red and green lights of the signal lights in the target area.

[0035] According to the third aspect provided by the present application, the present application provides an electronic device, including: a memory and a processor; the memory and the processor are coupled; the memory is used to store instructions executable by the processor; when the processor executes the instructions, it executes the traffic flow determination method as in the first aspect, and any possible implementation thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The drawings herein are incorporated into and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application, and do not constitute an improper limitation on the present application.

[0037] Figure 1 A method for determining traffic flow according to an exemplary embodiment is shown as a flow chart. Figure 1 ; Figure 2 A method for determining traffic flow according to an exemplary embodiment is shown as a flow chart. Figure 2 ; Figure 3 A method for determining traffic flow according to an exemplary embodiment is shown as a flow chart. Figure 3 ; Figure 4 A method for determining traffic flow according to an exemplary embodiment is shown as a flow chart. Figure 4 ; Figure 5 A method for determining traffic flow according to an exemplary embodiment is shown as a flow chart. Figure 5 ; Figure 6The figure is a structural diagram of a device for determining traffic flow according to an exemplary embodiment. DETAILED DESCRIPTION

[0038] In order to enable ordinary people in the art to better understand the technical solutions of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0039] It should be noted that the terms "first," "second," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.

[0040] In the embodiments of this application, words such as "exemplary," "for example," or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary," "for example," or "for example" in the embodiments of this application should not be interpreted as being preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary," "for example," or "for example" is intended to present the relevant concepts in a concrete manner.

[0041] The technical solutions in the embodiments of the present application will be described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments.

[0042] like Figure 1 As shown, an embodiment of the present application provides a method for determining traffic flow, which includes: S101-S103.

[0043] S101: Acquire a plurality of traffic flow data of a target area within a first time period.

[0044] The target area refers to a range that serves as the object of traffic flow analysis and may include one road or multiple roads.

[0045] The multiple traffic flow data correspond to different data sources and include traffic monitoring video data, vehicle positioning data, social media data, and weather data.

[0046] In one possible implementation, traffic flow data from multiple data sources is obtained, and the obtained traffic flow data is integrated into a multi-source traffic flow dataset. The multi-source traffic flow dataset is preprocessed to obtain a standard dataset.

[0047] Furthermore, for the traffic data at each moment of each data source in the standard data set, a plurality of different types of traffic data included at each moment are fused to obtain the traffic flow data corresponding to the data source, thereby obtaining a plurality of traffic flow data.

[0048] Specifically, for traffic data of a data type, the weight of each type of traffic data in the traffic data of the data type is determined, and based on the weight of each type of traffic data, multiple different types of traffic data in the data type are fused to obtain traffic flow data corresponding to the data source.

[0049] In one example, traffic monitoring video data includes vehicle location data and vehicle density data. Based on the weights of the vehicle location data and vehicle density data, the vehicle location data and vehicle density data in the traffic monitoring video data are fused according to the weight ratio to obtain traffic flow data corresponding to the traffic monitoring video data.

[0050] S102: Perform feature fusion on multiple traffic flow data to obtain a traffic feature dataset of the target area, and perform time series analysis on the traffic feature dataset to obtain traffic flow change information of the target area.

[0051] The feature fusion refers to fusing feature data from multiple traffic flow data into one data set, and the traffic feature data set includes all feature-fused traffic flow data within a first time period.

[0052] In one possible implementation, a feature extraction function is used to extract features from multiple traffic flow data sets to obtain a set of feature parameters corresponding to each traffic flow data set. The set of feature parameters includes traffic flow feature data at multiple moments in a first time period. The traffic feature data at the same moment in the feature parameter sets corresponding to the multiple traffic flow data sets are fused to obtain a traffic feature dataset for the target area. The traffic feature dataset includes both linear feature data and nonlinear feature data.

[0053] In another possible implementation, the linear feature data is input into a linear data analysis model, and the linear feature data is analyzed by the linear data analysis model to obtain first change information of the traffic flow in the target area.

[0054] The first change information includes a change trend of the traffic flow in the target area under the influence of the linear characteristic data.

[0055] The nonlinear characteristic data is input into a nonlinear data analysis model, and the nonlinear characteristic data is analyzed by the nonlinear data analysis model to obtain second change information of the traffic flow in the target area. The first change information and the second change information are data-fused to obtain change information of the traffic flow in the target area.

[0056] The second change information includes a change trend of the traffic flow in the target area under the influence of nonlinear data.

[0057] The traffic flow change information of the target area includes the fluctuation, repetition and long-term change trend of traffic flow data caused by changes in characteristic data.

[0058] S103: Based on the change information of the traffic flow in the target area, predict the traffic flow in the target area in a second time period.

[0059] The second time period refers to a time period after the first time period.

[0060] In a possible implementation, information on changes in traffic flow in the target area is input into a traffic flow prediction model to obtain predicted traffic flow data for the target area within a second time period.

[0061] Specifically, the vehicle flow speed, density, location change characteristics, weather condition characteristics, social media characteristics at the latest time point in the traffic flow change information of the target area and the historical traffic flow characteristics at the same time point as the latest time point in the past dates are input into the traffic flow prediction model, and the traffic flow prediction model outputs the predicted traffic flow data at the initial time point of the second time period .

[0062] in, represents a time point after the preset time point t in the first time period, for example It can be 1 minute, then Indicates the traffic flow in the target area 1 minute after the preset time point t in the first time period.

[0063] In another possible embodiment, the above The historical traffic data features of the next time window are input into the traffic flow prediction model, so that the traffic flow prediction model continues to iterate and finally outputs the traffic flow prediction data of the target area in the second time period.

[0064] In another possible embodiment, route planning is performed based on the predicted traffic flow data of the target area in the second time period, and a target driving route in the second time period is determined by a graph theory algorithm with the goal of minimizing the overall travel time.

[0065] The target driving route includes a route with the shortest driving time and / or a route with the least vehicles among multiple routes within the target area.

[0066] The target route plan is sent to the navigation system of the target vehicle that is about to enter the target area, guiding the target vehicle to pass through the target area in the shortest time. The target route satisfies the following formula 1: Formula 1.

[0067] in, represents the best route determined based on the predicted traffic flow, represents the estimated travel time of the i-th route through the target area, Is a positive integer.

[0068] In another possible embodiment, based on the predicted traffic flow data of the target area in the second time period and the existing signal control scheme, the traffic light display duration of the traffic lights in the target area in the second time period is adjusted using a traffic signal control optimization model, and the traffic signal control optimization model satisfies the following formula 2: Formula 2.

[0069] in, represents the traffic efficiency under the jth signal light adjustment plan, Is a positive integer.

[0070] It should be noted that in actual usage scenarios, the predicted traffic flow data of the target area in the second time period is generally expressed as the trend of the traffic flow in the target area changing with the characteristic data in the second time period. For example: for every 1 mm increase in rainfall in the target area, the number of vehicles passing through the target area per minute decreases by 2.3.

[0071] according to Figure 1 By using the technical means in the present application, a traffic characteristic data set of the target area is obtained by obtaining multiple traffic flow data of the target area in the first time period, and performing feature fusion on the multiple traffic flow data. The traffic characteristic data set is subjected to time series analysis to obtain traffic flow change information of the target area, and the predicted traffic flow data of the target area in the second time period is determined based on the change information. By fusing data from multiple data sources and the different feature data corresponding to the data sources, the changing trend of traffic flow under the influence of different feature data is analyzed. The traffic flow data of the target area in the future period can be accurately predicted, thereby alleviating traffic pressure and improving travel efficiency.

[0072] In one embodiment, Figure 2 As shown, the above method S101 also includes: S201-S203.

[0073] S201: Obtain traffic flow data from multiple data sources and integrate them into a multi-source traffic data set.

[0074] The specific methods for obtaining these four types of traffic flow data are described in detail below.

[0075] 1-1. Traffic monitoring video data.

[0076] The traffic monitoring video data refers to the traffic monitoring video data captured by the road monitoring camera in the target area during the first time period, and is used to obtain the vehicle flow speed and density in the target area during the first time period.

[0077] In one possible implementation, a traffic management system collects traffic monitoring video data V captured by surveillance cameras on roads in a target area during a first time period. At each time point t, video analysis techniques are used to analyze the traffic monitoring video data V, extracting information about the vehicle flow speed and vehicle density on the roads in the target area. The vehicle flow speed and vehicle density information are integrated to form a real-time traffic monitoring dataset. The traffic monitoring dataset satisfies the following formula 3: Formula 3.

[0078] in, Represents the vehicle flow information of the i-th monitoring point at time point t, where t, i, and n are all positive integers.

[0079] 1-2. Vehicle positioning data.

[0080] The vehicle positioning data refers to the GPS positioning data of all vehicles in the target area. The vehicle GPS signal data is used to obtain the position change information of the vehicles in the target area within a first time period.

[0081] In one possible implementation, a traffic management system collects GPS signals from vehicles traveling within the target area at each time point t. The GPS signals include the vehicle's location coordinates and timestamp. The collected GPS signals from vehicles within the target area are integrated to form a vehicle GPS dataset. The vehicle GPS dataset satisfies the following formula 4: Formula 4.

[0082] in, Represents the GPS signal information of the i-th vehicle at time point t.

[0083] 1-3. Social media data.

[0084] Among them, social media data refers to public posts in social media related to the traffic conditions in the target area.

[0085] In one possible implementation, by searching for keywords in the target social media and calling the social media search API, public posts related to traffic conditions in the target area are collected at each time point t within the first time period of the target area. Text analysis techniques are used to extract information from the collected public posts that affects traffic conditions in the target area, and this information is integrated to form a social media dataset. The social media dataset satisfies the following formula 5: Formula 5.

[0086] in, represents the i-th traffic-related social media post information at time point t.

[0087] 1-4. Weather data.

[0088] The weather data is used to obtain weather conditions of the target area within a first time period.

[0089] In one possible implementation, weather information data at each time point t in a first time period of a target area is obtained through a meteorological information system, and the obtained information data is integrated to obtain a weather information dataset. The weather information dataset satisfies the following formula 6: Formula 6.

[0090] in, Represents the weather information at time point t.

[0091] 1-5. Historical traffic flow data.

[0092] The historical traffic flow data refers to the traffic flow records in the target area during a certain period of time in the past that are the same as each time point t in the first time period.

[0093] In one possible implementation, a traffic management system collects and aggregates traffic flow records at each time point t within a first time period in the target area over a certain period of time in the past, including the number of vehicles passing through the target area and the average speed of vehicles passing through the target area, and integrates them into a historical traffic flow dataset. The historical traffic flow dataset satisfies the following formula 7: Formula 7.

[0094] in, Represents the historical traffic flow information at time point t.

[0095] Furthermore, the real-time traffic monitoring video data, vehicle GPS signal data, social media data, weather information data, and historical traffic flow data collected and integrated at time point t in the first time period of the target area are integrated to form a multi-source traffic dataset. The multi-source traffic dataset satisfies the following formula 8: Formula 8.

[0096] in, represents the traffic surveillance video data at time point t, represents the vehicle positioning signal data at time point t, represents social media data, represents the weather data at time point t, Represents the historical traffic data at time point t.

[0097] S202: Preprocess the multi-source traffic dataset to obtain a standard dataset.

[0098] Among them, preprocessing operations include data cleaning, data standardization, and data synchronization.

[0099] In one possible implementation, data cleaning, data standardization, and data synchronization are performed on multi-source traffic datasets to obtain a standard dataset.

[0100] Below, the data cleaning, data standardization, and data synchronization operations are explained in detail.

[0101] 2-1. Data cleaning.

[0102] Among them, data cleaning is used to clean multiple traffic flow data to remove inaccurate, incomplete or data irrelevant to the traffic conditions in the target area during the first time period.

[0103] In one possible implementation, an automated script is used to analyze each sub-dataset in the multi-source traffic dataset. , perform data identification and delete erroneous data points .

[0104] Among them, erroneous data points refer to data points that are missing key information or have incorrect formats.

[0105] Specifically, for the dataset Each data point in , the quality of data points is evaluated by a data quality evaluation function that comprehensively considers data integrity, format correctness, and content relevance. Data quality evaluation value Satisfy the following formula nine: Formula nine.

[0106] in, represents the data integrity weight, represents the format correctness weight, represents the content relevance weight, and , represents the data integrity function, represents the well-formedness function, Represents the content relevance function.

[0107] in, satisfy .

[0108] in, 、 and The value range of is [0,1], which means the degree of satisfaction ranges from 0 when it is completely unsatisfied to 1 when it is completely satisfied. For example, when the data point When the data integrity condition is not met at all, The value is 0; when the data point When the data integrity conditions are fully met, The value is 1.

[0109] When the data point Data quality assessment value Greater than or equal to the predetermined threshold , the data point is retained in the dataset.

[0110] When the data point Data quality assessment value Less than the predetermined threshold In the case of , the data point is deleted from the multi-source traffic dataset as an erroneous data point. The multi-source traffic dataset after deleting the erroneous data point is used as the first dataset.

[0111] 2-2. Data standardization.

[0112] Data standardization is used to standardize the data in the first data set and convert data from different data sources into the same format.

[0113] In one possible implementation, a unified data model is defined ,The model specifies a unified format for data in multi-source datasets, including the ,timestamp format , Position coordinate format , speed unit and text content standards . Define the transformation function to transform the data points The data is converted according to the unified data model M, and the converted data points are integrated to obtain a standardized data set.

[0114] It should be noted that for location data If the location data in the multi-source traffic dataset is expressed in latitude and longitude, and the unified format requires the use of the UTM coordinate system, a conversion function is used to convert the original location data into a coordinate system. The converted location information satisfies the following formula 10: Formula 10.

[0115] in, is the converted location information, It is a conversion function from latitude and longitude to UTM coordinate system.

[0116] The first data set after data normalization is used as the second data set.

[0117] 2-3. Data synchronization.

[0118] Data synchronization refers to aligning the timestamps of time-sensitive data in the second dataset so that all data are based on a unified time reference system.

[0119] In one possible implementation, a timestamp synchronization algorithm is used to adjust the timestamps of the data points in the second dataset so that all data points are based on a unified time reference system. After data synchronization, a standard dataset is obtained.

[0120] Among them, the standard data set refers to the multi-source traffic data set that has undergone data cleaning, data standardization, and data synchronization.

[0121] Specifically, the original timestamp of the data point is adjusted according to the time deviation and transmission delay of the data source to obtain a new timestamp. The new timestamp satisfies the following formula 11: Formula 11.

[0122] in, Indicates the new timestamp, Represents the original timestamp, represents a fixed time deviation, Indicates the transmission delay, All are positive integers.

[0123] The fixed time offset can be determined by the data source. The transmission delay can be determined by estimating the data transmission time. For details, please refer to the existing technology and will not be described in detail here.

[0124] S203: Fusing different types of traffic data at each moment from each data source in the standard data set to obtain a plurality of traffic flow data.

[0125] In one possible implementation, for the traffic data at each moment of each data source in the standard data set, multiple different types of traffic data included at each moment are fused to obtain traffic flow data corresponding to the data source, thereby obtaining multiple traffic flow data.

[0126] It should be noted that one data source corresponds to one data type, and one data source includes multiple types of traffic data at each preset moment in the first time period.

[0127] Specifically, for the traffic data of a data source in the standard data set, the weight of each type of traffic data in the traffic data of the data source is determined. Based on the weight of each type of traffic data, multiple different types of traffic data in the data type are fused according to the weight ratio to obtain the traffic flow data corresponding to the data source.

[0128] In one embodiment, Figure 3 As shown, the above method S102 also includes S301-S302.

[0129] S301 , performing feature extraction on a plurality of traffic flow data respectively through a feature extraction function to obtain a feature parameter set corresponding to each traffic flow data.

[0130] In one possible implementation, based on the weight corresponding to each data source, traffic characteristic data at the same time in the characteristic parameter sets corresponding to multiple traffic flow data are fused to obtain a first characteristic data set of the target area; The first feature data set refers to a data set including fused feature data of each moment in a plurality of moments.

[0131] Among them, the feature types in traffic flow data include the flow speed of vehicles in the target area, the density of vehicles in the target area, the position changes of vehicles in the target area, descriptions of traffic conditions in the target area on social media, and weather conditions in the target area.

[0132] Specifically, the feature extraction function is used to extract the sub-dataset at each time t in the above standard data set. Data points in Perform feature extraction to obtain multiple traffic feature types. Through the feature extraction function, based on the weight of each feature type's impact on traffic flow, the feature types are fused into first feature data according to the weight ratio. The first feature data satisfies the following formula 12: Formula twelve.

[0133] in, represents the first feature data at time point t, Represents the vehicle flow speed and density characteristics, Indicates the position change characteristics of vehicles in the target area, Descriptive features of traffic conditions within the target area of social media relationships, Indicates the weather conditions characteristics of the target area, Represents the historical traffic flow data characteristics of the target area, represents the i-th set at time point t, 、 、 、 、 represents the weight parameter, Is a positive integer.

[0134] It should be noted that the weight parameter is used to balance the contribution of different features in the first feature data, and can also be regarded as the degree of influence of different features on traffic flow. The weight parameter can be set according to actual needs.

[0135] S302: Fusing traffic characteristic data at the same time in characteristic parameter sets corresponding to multiple traffic flow data to obtain a traffic characteristic data set of the target area.

[0136] In a possible implementation, a convolution operation is performed on the fused feature data at multiple moments included in the first feature dataset to obtain a traffic feature dataset of the target area.

[0137] Specifically, the traffic characteristic vector satisfies the following formula 13: Formula 13.

[0138] in, represents the traffic feature vector at the preset time point t, represents the parameters of the convolutional neural network, represents the activation function of the output layer of the convolutional neural network, Indicates the first feature data at time point t.

[0139] It should be noted that the parameters of the convolutional neural network can be obtained by testing the convolutional neural network, and the activation function can be set as required.

[0140] Furthermore, for each time point t within the first time period, the above steps S202-S203 are repeated to obtain the traffic feature vector at each time point t within the first time period. The traffic feature vectors at each preset time point t are integrated into a set to obtain a traffic feature dataset at each time point t within the first time period. The traffic feature dataset satisfies the following formula 14: Formula 14.

[0141] in, represents the traffic feature dataset at time point t in the first time period, represents multiple comprehensive feature vectors at time point t, represents the parameters of the convolutional neural network, Represents the activation function of the output layer of the convolutional neural network.

[0142] In one embodiment, Figure 4 As shown, the above S102 also includes S401-S402.

[0143] S401: Inputting linear feature data into a linear data analysis model, analyzing the linear feature data through the linear data analysis model, and obtaining first change information of traffic flow in a target area.

[0144] In one possible implementation, a time series analysis is performed on the traffic characteristic dataset using a linear data analysis model to analyze the relationship between traffic flow changes and time. By combining the autoregressive, differencing, and moving average components of the linear data analysis model, the serialized traffic characteristic dataset is analyzed to determine the periodicity and trend of traffic flow changes, which are output as the first change information.

[0145] The linear data analysis model satisfies the following formula: Formula 15.

[0146] in, represents the autoregressive part, represents the differential part, Represents the moving average part Specifically, a longer time window (e.g., a week or a month) is set and a sliding average is performed on the traffic characteristic dataset. After filtering out daily fluctuations, a smooth curve is obtained. This smooth curve can reflect the long-term trend of overall traffic flow over time, that is, whether traffic flow maintains an upward or downward trend over time. A unit root test or stationarity test is performed on the sliding average result. If it is not stationary, it indicates that the traffic flow has a clear trend over time.

[0147] In one example, after removing the long-term trend, the remaining series data is used to plot the autocorrelation function (ACF) or partial autocorrelation function (PACF). If significant peaks in traffic flow occur at fixed time intervals, such as 24 hours or 168 hours, this indicates that traffic flow repeats within these fixed time periods. For example, traffic flow in the target area peaks at 9:00 AM every Monday.

[0148] In another example, even after removing long-term and cyclical trends, traffic flow is still affected by holidays, temperature, seasonal changes, and other factors. Using the seasonal-trend-residual decomposition method, the series after removing long-term and cyclical trends is grouped into hourly or daily time units, and traffic flow characteristics are calculated at each time point. For example, if traffic at 5 p.m. is 30% higher than usual, this is the seasonal component.

[0149] After confirming the long-term trend, cyclical trend and seasonal component, the linear data are subjected to first-order difference and cyclical difference (such as 24-hour difference or 168-hour difference) to remove the cyclical trend and make the remaining series stationary.

[0150] S402: Input the nonlinear characteristic data into a nonlinear data analysis model, analyze the nonlinear characteristic data using the nonlinear data analysis model, and obtain second change information of the traffic flow in the target area.

[0151] Among them, the nonlinear data analysis model is used to analyze the traffic flow changes of the nonlinear data in the traffic characteristic data set of the target area.

[0152] In a possible implementation, the nonlinear characteristic data is analyzed by a nonlinear data analysis model to obtain the second change information of the traffic flow in the target area.

[0153] Specifically, nonlinear data is input into a nonlinear data analysis model, and the gating mechanism of the nonlinear data analysis model is used to capture the long-term dependencies in the time series data and learn the relevant features in the traffic feature dataset. The nonlinear data analysis model satisfies the following formula 16: Formula 16.

[0154] in, and represents the weight and bias parameters of the nonlinear data analysis model, represents the activation function, Represents the combination of the hidden state at the previous time step and the input feature vector at the current time step.

[0155] The nonlinear data analysis model includes an input gate, a forget gate, and an output gate. The input gate determines which feature data enters the nonlinear data analysis model, the forget gate determines which feature information to retain or discard, and the output gate outputs data to the next layer based on the current state and new features.

[0156] In another possible implementation, the first difference is input into the input gate of the nonlinear data analysis model. Through the collaborative mechanism between the input gate, the forget gate, and the output gate, the noise data in the first difference is automatically filtered out, highlighting the long-term dependence of traffic flow data on time.

[0157] Furthermore, the error between the model's predicted values and historical traffic flow data is used through time backpropagation to iteratively adjust all gating units and state update parameters in the nonlinear analysis model. When the error is less than a preset threshold, the nonlinear analysis model accurately captures the short-term fluctuations, cyclical recurrences, and long-term trends in traffic flow as a function of nonlinear data. This information is then integrated into the output of traffic flow changes within the target area.

[0158] according to Figure 4 Using the technical means in this paper, this application uses a linear data analysis model to analyze linear feature data to obtain first information on changes in traffic flow in the target area, and uses a nonlinear data analysis model to analyze nonlinear feature data to obtain second information on changes in traffic flow in the target area. By analyzing these two data types, we comprehensively analyze the changes in traffic flow in the target area as affected by both linear and nonlinear data, providing reliable technical support for subsequent prediction of traffic flow data.

[0159] according to Figure 4 Using the technical means in this paper, this application uses a linear data analysis model to analyze linear feature data to obtain first information on changes in traffic flow in the target area, and uses a nonlinear data analysis model to analyze nonlinear feature data to obtain second information on changes in traffic flow in the target area. By analyzing these two data types, we comprehensively analyze the changes in traffic flow in the target area as affected by both linear and nonlinear data, providing reliable technical support for subsequent prediction of traffic flow data.

[0160] In one example, Figure 5 As shown, the method provided in the embodiment of the present application also includes S501.

[0161] S501: Construct a traffic flow prediction model based on traffic flow change information in a target area.

[0162] In one possible implementation, characteristic variables affecting traffic flow are determined based on traffic flow change information in a target area, and a traffic flow prediction model is constructed based on the characteristic variables affecting traffic flow.

[0163] Among them, the characteristic variables that affect traffic flow include vehicle flow speed, density, location change characteristics, weather condition characteristics, social media characteristics, and historical traffic flow characteristics.

[0164] Specifically, an initial traffic flow prediction model is constructed based on the above-mentioned characteristic variables affecting traffic flow. The initial traffic flow prediction model satisfies the following formula 17: Formula 17.

[0165] in, represents the predicted traffic flow at the preset time point t, Represents the i-th feature in the traffic feature dataset, including features ranging from vehicle flow speed, density to location changes, represents the features related to the jth weather condition, including temperature and rainfall, represents the kth social media metric, including the number of traffic-related posts or sentiment analysis results, Indicates the Features related to historical traffic flow data, represents the model parameters, represents the error term.

[0166] Specifically, a fixed time sliding window is set to compare the traffic characteristic data at the preset time point t, the weather condition dataset, the social media dataset, the historical traffic flow dataset and the actual traffic flow at the time point t in the historical traffic flow data. The time dimension of the dataset is aligned to form the training dataset.

[0167] In another possible implementation, a training dataset is input into an initial traffic flow prediction model. When the deviation between a predicted result corresponding to the target historical traffic flow data and an actual result corresponding to the target historical traffic flow data is greater than a preset threshold, the initial traffic flow prediction model is iteratively trained until the difference between the predicted result and the actual result of the trained traffic flow prediction model is less than or equal to the preset threshold. The trained traffic flow prediction model is then used as the traffic flow prediction model.

[0168] Specifically, the training dataset is input into the initial traffic flow prediction model, which outputs the predicted traffic flow. The RMSE and MAPE loss functions are introduced to train the traffic flow prediction model. The RMSE loss function is used to calculate the root mean square error (RMSE) between the predicted traffic flow output by the initial traffic flow prediction model and the actual traffic flow. The MAPE loss function is used to calculate the mean absolute deviation (MAD) between the predicted traffic flow output by the initial traffic flow prediction model and the actual traffic flow. If the RMSE is less than or equal to 30% of the data standard deviation and the MAD is less than 10%, the initial traffic flow prediction model is considered successfully trained and a traffic flow prediction model is obtained.

[0169] According to the training data set, the loss function RMSE is introduced to calculate the root mean square error of the initial traffic flow prediction model. The root mean square error satisfies the following formula 18: Formula 18.

[0170] in, represents the actual traffic flow at time point t, represents the predicted traffic flow at time point t, represents the root mean square error, and N represents the training data set.

[0171] It should be noted that the root mean square error is used to quantify the average absolute deviation between the predicted value and the true value, and intuitively reflects the prediction accuracy of the model.

[0172] The loss function MAPE is introduced to calculate the mean absolute percentage error of the model based on the training data set. The mean absolute percentage error satisfies the following formula 19: Formula 19.

[0173] in, represents the actual traffic flow at time point t, represents the predicted traffic flow at time point t, represents the mean absolute percentage error, and N represents the training data set.

[0174] Furthermore, if the collinearity of the independent variables in the training data set of the initial traffic flow prediction model is not strong, the least squares estimation is used. 、 If there is obvious multicollinearity between the independent variables, a ridge regularization term is added to the loss function. Satisfy the following formula twenty: Formula 20.

[0175] in, represents the response variable vector, represents the regression coefficient vector to be estimated, represents the regularization parameter for ridge regression.

[0176] According to the above formula 16, we get its analytical solution: Satisfy the following formula 21: Formula 21.

[0177] in, represents the analytical solution of the normal equation of coefficients, represents the matrix X, represents the response variable vector, represents the identity matrix of the same dimension, represents the regularization parameter for ridge regression.

[0178] Obtain the parameter vector by solving the normal equations or gradient descent method , and calculate the standard error of each coefficient , use t statistics to test whether the standard error of each coefficient is significant. Satisfy the following formula 22: Formula 22.

[0179] in, represents the estimated value of the i-th regression coefficient, represents the standard error of the i-th coefficient, Represents statistics.

[0180] Calculating the variance inflation factor , Satisfy the following formula 23: Formula 23.

[0181] in, Indicates that the The independent variable has no effect on other The coefficient of determination obtained when regressing the independent variables is represents the variance inflation factor.

[0182] The variance inflation factor is used to measure the variance of the independent variable The degree of multicollinearity between the independent variables.

[0183] Between variables In the case of , the model parameters of the initial prediction model are adjusted by increasing the regularization strength or eliminating variables.

[0184] After the initial traffic flow prediction model is trained according to the above method, the model is evaluated and the determination coefficient is calculated on the validation data set. The determination coefficient satisfies the following formula 24: Formula 24.

[0185] in, represents the coefficient of determination, represents the actual traffic flow at time point t, represents the predicted traffic flow at time point t.

[0186] It should be noted that the coefficient of determination is used to measure the ability of the model to explain the variation of the dependent variable. The closer the value of is to 1, the stronger the model's ability to explain data variation. When >0.9, the ability of the traffic flow prediction model to explain data variation meets the required requirements.

[0187] It should be noted that the mean absolute percentage error refers to the relative error between the predicted traffic flow and the actual traffic flow measured in percentage.

[0188] When RMSE is less than or equal to 30% of the data standard deviation and MAPE is less than or equal to 10%, the traffic flow prediction model is obtained.

[0189] It should be noted that the thresholds of RMSE and MAPE here are set according to the specific requirements of the traffic flow prediction model. In actual use, the thresholds of RMSE and MAPE can be adjusted according to different prediction needs to obtain a suitable traffic flow prediction model.

[0190] Through the description of the above implementation methods, technical personnel in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete the full classification or partial functions described above.

[0191] according to Figure 5 This application uses the technical means in the previous section to train an initial traffic flow prediction model. When the difference between the predicted results of the trained traffic flow prediction model and the actual results is less than or equal to a preset threshold, the trained traffic flow prediction model is used as the traffic flow prediction model. This provides technical support for subsequent traffic flow predictions, making the prediction results more accurate.

[0192] like Figure 6 As shown in FIG, the present application provides a structural diagram of a device for determining traffic flow. Figure 6 As shown, the traffic flow determination device can be used to perform Figure 1 、 Figure 2 、 Figure 3 、 Figure 4 、 Figure 5 The traffic flow determination method shown in the figure, the traffic flow determination device 60 includes: an acquisition module 601 and a determination module 602.

[0193] An acquisition module 601 is configured to acquire a plurality of traffic flow data of a target area within a first time period; Determination module 602, for performing feature fusion on multiple traffic flow data to obtain a traffic feature dataset of a target area, and performing time series analysis on the traffic flow dataset to obtain traffic flow change information of the target area; The determination module 602 is further configured to determine traffic flow prediction data of the target area within a second time period based on the change information of the traffic flow in the target area.

[0194] In one possible approach, module 602 is determined, the traffic feature data set includes linear feature data and nonlinear feature data; a time series analysis is performed on the traffic flow data set to obtain change information of the traffic flow in the target area, including: inputting the linear feature data into a linear data analysis model, analyzing the linear feature data through the linear data analysis model, and obtaining first change information of the traffic flow in the target area; inputting the nonlinear feature data into a nonlinear data analysis model, analyzing the nonlinear feature data through the nonlinear data analysis model, and obtaining second change information of the traffic flow in the target area; and performing data fusion on the first change information and the second change information to obtain change information of the traffic flow in the target area.

[0195] In one possible approach, the determination module 602 determines the predicted traffic flow data of the target area in the second time period based on the change information of the traffic flow in the target area, including: inputting the change information of the traffic flow in the target area into the traffic flow prediction model, and the predicted traffic flow data of the target area in the second time period.

[0196] In one possible manner, in the determination module 602, the traffic flow prediction model is determined in the following manner: the target historical traffic flow data is input into the initial traffic flow prediction model to obtain a prediction result corresponding to the target historical traffic flow data; when the deviation between the prediction result corresponding to the target historical traffic flow data and the actual result corresponding to the target historical traffic flow data is greater than a preset threshold, the initial traffic flow prediction model is iteratively trained until the difference between the prediction result of the trained traffic flow prediction model and the actual result is less than or equal to the preset threshold, and the trained traffic flow prediction model is used as the traffic flow prediction model.

[0197] In one possible approach, the determination module 602 performs feature fusion on multiple traffic flow data to obtain a traffic feature data set of the target area, including: performing feature extraction on the multiple traffic flow data using a feature extraction function to obtain a feature parameter set corresponding to each traffic flow data; the feature parameter set includes traffic flow feature data at multiple moments in a first time period; and fusing the traffic feature data at the same moment in the feature parameter set corresponding to the multiple traffic flow data to obtain the traffic feature data set of the target area.

[0198] In one possible approach, the determination module 602 fuses the traffic feature data at the same time in the feature parameter sets corresponding to the multiple traffic flow data to obtain a traffic feature data set of the target area; based on the weight corresponding to each data source, the traffic feature data at the same time in the feature parameter sets corresponding to the multiple traffic flow data are fused to obtain a first feature data set of the target area; the first feature data set includes the fused feature data at each of the multiple moments, and a convolution operation is performed on the fused feature data at the multiple moments included in the first feature data set to obtain a traffic feature data set of the target area.

[0199] In one possible approach, the determination module 602 obtains multiple traffic flow data of the target area within the first time period, including: obtaining traffic information of each data source within the first time period; the traffic information includes traffic data at multiple moments, and the traffic data at each moment includes multiple different types of traffic data; for the traffic data at each moment in each data source, the multiple different types of traffic data included at the moment are integrated to obtain traffic flow data corresponding to the data source, so as to obtain multiple traffic flow data.

[0200] In one possible approach, the determination module 602 fuses multiple different types of traffic data included at a moment to obtain traffic flow data corresponding to the data source, including: determining the weight of each type of traffic data; and fusing multiple different types of traffic data based on the weight of each type of traffic data to obtain traffic flow data corresponding to the data source.

[0201] In one possible embodiment, the method further includes determining module 602 and performing route planning based on traffic flow prediction data of the target area within the second time period to determine a target driving route within the second time period; the target driving route includes a route with the shortest driving time and / or a route with the least vehicles among multiple routes within the target area; and sending the target driving route to the target vehicle.

[0202] In one possible manner, the determination module 602 further includes: determining the traffic light control information of the target area within the second time period based on the traffic flow of the target area within the second time period; the traffic light control information is used to control the display duration of the red and green lights of the traffic lights in the target area.

[0203] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of modules or units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another device, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0204] The units described as separate components may or may not be physically separate, and the components shown as units may be one physical unit or multiple physical units, that is, they may be located in one place or distributed in multiple places. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0205] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0206] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiment of the present application, or the part that contributes to the prior art, or the entire classification part or part of the technical solution can be embodied in the form of a software product. The software product is stored in a storage medium and includes a number of instructions for enabling a device (which can be a single-chip microcomputer, chip, etc.) or a processor to execute the entire classification part or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, ROM, RAM, disk or optical disk, etc., various media that can store program code.

[0207] The above are only specific embodiments of the present application, but the scope of protection of the present application is not limited thereto. Any changes or replacements within the technical scope disclosed in this application should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for determining traffic flow, characterized in that: The method comprises: Acquire a plurality of traffic flow data of a target area within a first time period; the plurality of traffic flow data corresponds to different data sources; Performing feature fusion on the plurality of traffic flow data to obtain a traffic feature dataset of the target area, and performing time series analysis on the traffic feature dataset to obtain change information of the traffic flow in the target area; Based on the change information of the traffic flow of the target area, the predicted traffic flow data of the target area in a second time period is determined; the second time period is located after the first time period.

2. The determination method according to claim 1, characterized in that The traffic characteristic data set includes linear characteristic data and nonlinear characteristic data; performing time series analysis on the traffic characteristic data set to obtain change information of traffic flow in the target area includes: Inputting the linear feature data into a linear data analysis model, and analyzing the linear feature data using the linear data analysis model to obtain first change information of the traffic flow in the target area; inputting the nonlinear characteristic data into a nonlinear data analysis model, and analyzing the nonlinear characteristic data using the nonlinear data analysis model to obtain second change information of the traffic flow in the target area; Data fusion is performed on the first change information and the second change information to obtain change information of the traffic flow in the target area.

3. The determination method according to claim 1 or 2, characterized in that: The determining, based on the traffic flow change information of the target area, predicted traffic flow data of the target area in the second time period includes: Inputting the traffic flow change information of the target area into the traffic flow prediction model, the predicted traffic flow data of the target area in the second time period.

4. The determination method according to claim 3, characterized in that: The traffic flow prediction model is determined by: Inputting the target historical traffic flow data into the initial traffic flow prediction model to obtain a prediction result corresponding to the target historical traffic flow data; When the deviation between the predicted result corresponding to the target historical traffic flow data and the actual result corresponding to the target historical traffic flow data is greater than a preset threshold, the initial traffic flow prediction model is iteratively trained until the difference between the predicted result of the trained traffic flow prediction model and the actual result is less than or equal to the preset threshold, and the trained traffic flow prediction model is used as the traffic flow prediction model.

5. The determination method according to claim 1 or 2, characterized in that: The step of performing feature fusion on the plurality of traffic flow data to obtain a traffic feature dataset of the target area includes: Performing feature extraction on the plurality of traffic flow data using a feature extraction function to obtain a feature parameter set corresponding to each traffic flow data; the feature parameter set includes traffic flow feature data at a plurality of moments in the first time period; The traffic characteristic data at the same time in the characteristic parameter sets corresponding to the multiple traffic flow data are fused to obtain a traffic characteristic data set of the target area.

6. The determination method according to claim 5, characterized in that: fusing the traffic characteristic data at the same time in the characteristic parameter sets corresponding to the plurality of traffic flow data to obtain a traffic characteristic data set of the target area; Based on the weight corresponding to each data source, the traffic feature data at the same time in the feature parameter sets corresponding to the multiple traffic flow data are fused to obtain a first feature data set of the target area; the first feature data set includes fused feature data at each time of the multiple time points; A convolution operation is performed on the fused feature data at multiple moments included in the first feature data set to obtain a traffic feature data set of the target area.

7. The determination method according to claim 1 or 2, characterized in that: The acquiring of a plurality of traffic flow data of the target area within the first time period includes: Obtaining traffic information of each data source within the first time period; the traffic information includes traffic data at multiple times, and the traffic data at each time includes multiple different types of traffic data; For the traffic data at each moment in each data source, a plurality of different types of traffic data included in the moment are fused to obtain the traffic flow data corresponding to the data source, so as to obtain the plurality of traffic flow data.

8. The determination method according to claim 7, characterized in that: The fusing of the multiple different types of traffic data included at the moment to obtain the traffic flow data corresponding to the data source includes: Determine the weight of each type of traffic data; Based on the weight of each type of traffic data, the multiple different types of traffic data are fused to obtain traffic flow data corresponding to the data source.

9. The determination method according to claim 1 or 2, characterized in that: The method further comprises: performing route planning based on the predicted traffic flow data of the target area during the second time period, and determining a target driving route during the second time period; the target driving route comprising a route with the shortest driving time and / or a route with the least vehicles among a plurality of routes within the target area; The target driving route is sent to the target vehicle.

10. The determination method according to claim 1 or 2, characterized in that: The method further comprises: Based on the predicted traffic flow data of the target area in the second time period, signal light control information of the target area in the second time period is determined; the signal light control information is used to control the display duration of the red and green lights of the signal lights in the target area.

11. A device for determining traffic flow, characterized in that: The device comprises: An acquisition module, configured to acquire a plurality of traffic flow data of a target area within a first time period; a determination module, configured to perform feature fusion on the plurality of traffic flow data to obtain a traffic feature dataset of the target area, and perform time series analysis on the traffic feature dataset to obtain change information of the traffic flow of the target area; The determination module is further configured to determine traffic flow prediction data of the target area within a second time period based on the change information of the traffic flow in the target area.

12. An electronic device, characterized in that: include: memory and processor; The memory is coupled to the processor; The memory is used to store instructions executable by the processor; When the processor executes the instructions, the determination method according to any one of claims 1 to 10 is performed.

Images